The Efficiency Toolbox
December 19, 2012 |Estimated reading time: 8 minutes
Editor's note: This article originally appeared in the November 2012 issue of The PCB Magazine.
Throughout the years, various manufacturing techniques have been accepted, bandied about or become the rule by which to run a factory. All of them have merit, but no one is the panacea for all efficiency drives. Those companies that have been set up as training vehicles or as consultancies for implementation have made a living of one sort or another. We have seen Lean manufacturing experts or Six Sigma experts, amongst many, who have flourished. But what should the efficiency toolbox consist of?
The key to all futures is the adherence to continuous improvement, also known as Kaizen. Whatever the technique chosen or rule bought into, none will work effectively unless continuous improvement is behind the choice. Merely using one of the techniques to attack a particular problem one time will not take a company forward. The decision to use efficiency techniques must be backed with a commitment to continue to use these techniques, and possibly others as well, forever.
Behind all of this lies a huge amount of common sense. None of the efficiency techniques is rocket science. All of them are methodical ways of approaching the concept of improving efficiency.
Before we even start to consider how to make a manufacturing system or a product more efficient, we should look at the whole chain of events. Today, this is commonly called logistics. It makes sense to cut out inefficiency when making something. None of us wants any waste. It also makes sense to build a product to the highest quality possible. Field returns will be fewer, warranty claims will be reduced and public, or consumer satisfaction will be maintained. The choice of parts; the product design; the bill of materials; the way it is assembled; the quality control; the way it is stored and packed; and finally, the way it is shipped to its final destination all carry efficiency burdens, but all are of different types. Therefore, none of the techniques that are popular today will solve all of the issues.
The two most popular efficiency methods today are Lean manufacturing, which is production-based, and Six Sigma, which focuses on product quality. It will pay us to understand a little more about each of them.
Six Sigma
This is a statistics-based regime. Sigma is the term used to define a standard deviation in mathematical statistics, which is a measure of diversity or variability in a statistical population, and shows how much variation there is from the average or mean. The picture below depicts a typical Gaussian curve that shows a spread of events/occurrences around a central axis and has a +/-3-sigma spread. Clearly, anything close to the central axis will be as close as possible to the required factors, be they quality or anything else; but, as soon as anything goes away from the centre, defects become possible. If the spread is six sigma, everything becomes much tighter and the risk of error reduces.
A one sigma system is not very efficient and the three sigma system shown above is not much better. Six sigma systems have a very small deviation indeed and if we translate the Gaussian curve into percentage errors we get the following table:
A three sigma system offers 66,807 defects per million opportunities, which is high by anyone’s standards. So, only 3.4 defects per million opportunities is an excellent target to aim for in product terms and it is achievable, but hard. Basically, this translates to do it right the first time. We must also ask where we should measure six sigma, in returns from the field or in the factory before the product is shipped. Ideally, we should measure just before the product is shipped. Measuring field returns masks factory inefficiencies.
What about waste or organisation in the factory? Six Sigma does not offer all the answers and we have to adopt methods such as Lean manufacturing to get at these inefficiencies.
Lean Manufacturing
A well-known definition of Lean manufacturing can be stated as follows: A systematic approach to identifying and eliminating waste (non-value-add activities) through continuous improvement by flowing the product at the pull of the customer in the pursuit of perfection. We can define many of the waste items (but by no means all) as follows, but the most important fact to acknowledge is that customers are not prepared to pay for waste.
- Overproduction: The classic build-for-stock situation in the hope that everything will be sold
- Waiting: Why wait for something to happen when all it does is cause expensive delays?
- WIP (work in progress): Most of us like the safety-blanket of having some partially-finished product ready for the next stage, but few know how much work-in-progress costs
- Processing waste: Many of us make things needlessly because they never get sold or made profitable. Everything we do should have a profit margin in mind
- Transportation: It is ecologically good to advertise that parts for a product come from local sources. The same thing applies to the food we eat. At one time, the lemmings amongst us rushed to offshore locations such as Asia to get product built supposedly cheaper. If the product is designed in Europe and its biggest market is in Europe, why build it in Asia and suffer huge transport costs and time penalties? After all, we are supposed to adopt the principles of time-to-market now so that delays are unacceptable
- Motion: If a product can be sequentially built, (i.e., built with the minimum number of turnovers) then the motion it goes through will be minimised
- Making defects: At one time, it was the norm to allow for defects to be made and then hope that only good product is allowed to be sold. Surely, it is far better to stop making defects
- Under-utilised staff: Nobody wants to be rushed off their feet for their working day but, equally, nobody wants to be bored stiff either. Staff should be consistently busy without being over-worked, but there should be no excess labour. We do not want to pay for people to sit around and do nothing and they do not want to feel under-utilised either.
We can now consider another method that could come under the Lean manufacturing umbrella. It is called Total Productivity Maintenance. It, and Lean, covers items such as the following.
- Unexpected breakdown: This usually applies to automated equipment, but could equally apply to a critical member of staff not being available at the required time. Normally, it relates to equipment failure outside of routine maintenance intervals. Most equipment manufacturers suggest that there should be at least 2,000 hours MTBF (mean time between failure) and, these days, this figure is usually 3,000 hours or more. Equally important is the related MTTR (mean time to repair) figure, which should be around two hours. It could be more if the necessary engineering involvement has to travel long distances
- Set-up and adjustment losses: If a product flows down a production line, but hesitates regularly because the equipment needs to be adjusted, or if it takes an inordinately long time to set up in the first place, there will be losses in productivity. Some companies just accept this as fait accompli
- Idling and minor stoppage time: Replenishing items used in the manufacturing process should not cause any delays. Equally, why does a line stop for no apparent reason? Even if the stoppage time is small, it should not happen in the first place
- Speed too low: This is so obvious that it is often overlooked. If a line can run faster, then make it do so. This extra line speed can then allow more product, or different products to be produced. If it is too low because productivity does not demand high-speed yet, then the decision was strategic for the future
- Defects and rework: Many companies have large rework squads aimed at capturing defects before products escape into the shipping department. Making it right the first time will eliminate expensive rework squads and the experienced staff can be released for other tasks. The only time a rework squad can be contemplated is when the product cost, complexity and volume dictate that all the products must be perfect. Some avionics and telecoms equipment is built to very exacting standards and has a very expensive BoM. If you can get rid of rework, then do
- Start-up yield losses: Some products seem easy to make and run at full speed as soon as they are put into production. Others are the opposite. Much of this is related to DFM (design for manufacture). If a product is designed to give the manufacturing group the lowest number of anxious moments possible, then it will usually be an efficient product.
How many times have you walked through a factory and seen piles of WIP or lines not running or people not working? All of these cost money, as do all of the things mentioned above.
We have looked at a structured approach to a couple of methods, but the main message here is that a common-sense attitude and the knowledge that the efficiency toolbox has many different tools in it work best of all. The last word to emphasize could be Kaizen, or continuous improvement. Kaizen requires the use of all the tools in the box, not just one. The common-sense, Kaizen approach requires us to:
- Standardise operations and activities
- Measure the standardised operation
- Gauge it against requirements
- Innovate to meet requirements and increase productivity
- Standardise the new improved operations and then go through the cycle again to continuously improve
How you do this involves all the tools in the toolbox. Wouldn’t it be good if we were all efficient in a common-sense way!
Peter Grundy is a manufacturing consultant and contract engineering resource working in the electronics industries of Europe and North America. His 30-year career includes experience in roles ranging from IT through die-casting, CNC machine tools, PCB drilling equipment, product management and production consulting, as well as NPI, Lean, OEE, factory planning and continuous improvement.